Cluster Of Workstations Research Articles

An Adaptive Parallel Pattern Based Design for Molecular Dynamic Simulation

In Parallel programming, a programmer needs to understand hardware environment, programming paradigm and primitives available in the programming language. Most of the time, parallel programmes are written for a specific architecture and cannot typically adapt to other architectures Particularly, programs written for shared memory architectures are unsuitable for distributed or hybrid architectures. This paper proposes Adaptive Design Pattern for Parallel Programming to improve adaptability, flexibility with achieving performance on different architectures. Molecular Dynamics (MD) simulation is required to scale to various architectures from simple machine to cluster of workstations. In this study, MD Simulation experimented using both pure benchmark code and code based on adaptive design patterns. Redesigned MD Simulation with Adaptive Design Pattern claims parallel efficiency from 56% to 90% for different number of processing elements used. The solution demonstrates adaptability to different architectures and scalability to use with large number of atoms and long duration simulation.

AUTOMATING BUILD OF EMBEDDED LINUX BINARIES IN A HIGH-PERFORMANCE COMPUTING CLUSTER

A method for automating the build process of the embedded Linux operating system binary images in a high-performance computing cluster based on continuous integration and continuous deployment technology is discussed. The method proposed can significantly reduce software design time, free up computing resources at local workstations and organize joint work on a team project. A practical example of setting up the software tools necessary to compile the boot-loader, embedded Linux kernel and root file system for BeagleBone Black device with an ARM Cortex A8 processor is provided. Basic requirements for the server and workstation software are formulated. The materials of the article can be used to organize system software design for Internet of Things devices effectively, as well as for educational purposes. Tools used to develop embedded operating systems based on Linux Yocto as well as GitLab source code management system and configuring of CI/CD feature are described. The build system environment proposed includes GitLab service installed on a high-performance virtual machine in a computing cluster and set of workstations connected to the local network. The project repository is physically located on the server and managed by GitLab service. Docker virtualization is used to isolate build process from the server operating system. Building of operating system image is started automatically using CI/CD runner script, every time changes to the source code of the project are made. Git protocol is used to modify project repository remotely. Intermediate compilation data is stored outside of the Docker sandbox to prevent data integrity between compilation events. This data includes a significant amount of third-party library source code downloaded from external Linux repositories on the Internet. Thus, rebuilding the project after making changes does not require a complete rebuild and reloading of library sources, and takes much less time. Requirements for the workplace of a development engineer are minimized because the main computing load is transferred to a remote server. The proposed method of building an image of the Linux Yocto operating system allows manage team to work on a project and significantly increases the speed of the project rebuild procedure, therefore, reduces the time spent on designing and debugging system software for embedded systems. The effectiveness of the proposed method was confirmed practically with sample project Linux Yocto for the BeagleBone Black minicomputer based on the ARM Cortex A8 processor. The article contains examples of settings taken from this project. The proposed method can be used to study design of system software for embedded systems. Handson training sessions are a good example of team-based software development in a time-constrained environment.

Estimating the overhead and coupling of scientific computing clusters

Since simulation became the third pillar of scientific research, several forms of computers have become available to drive computer aided simulations, and nowadays, clusters are the most popular type of computers supporting these tasks. For instance, cluster settings, such as the so-called supercomputers, cluster of workstations (COW), cluster of desktops (COD), and cluster of virtual machines (COV) have been considered in literature to embrace a variety of scientific applications. However, those scientific applications categorized as high-performance computing (HPC) are conceptually restricted to be addressed only by supercomputers. In this aspect, we introduce the notions of cluster overhead and cluster coupling to assess the capacity of non-HPC systems to handle HPC applications. We also compare the cluster overhead with an existing measure of overhead in computing systems, the total parallel overhead, to explain the correctness of our methodology. The evaluation of capacity considers the seven dwarfs of scientific computing, which are well-known, scientific computing building blocks considered in the development of HPC applications. The evaluation of these building blocks provides insights regarding the strengths and weaknesses of non-HPC systems to deal with future HPC applications developed with one or a combination of these algorithmic building blocks.

WITHDRAWN: Enterprise financial investment platform simulation based on FPGA-based high-performance algorithm

Abstract Comprehensive economic evaluation of investment decision-making process is an urgent need to ensure the accuracy. System dynamics is illustrated as an example of stage performances as a tool artist and is associated with local projects to set up models. Proposed model is mainly based on the model, the simulation of the actual project, focused on the financial evaluation and analysis of the project, the profitability of a comprehensive and dynamic understanding is, get the economic rationality of ability to pay off debt and project it can further the sensitivity of the system is detected to determine the key factors that affect the viability of the model. The most sensitive factor is, in the freight and volume, the model is effective and feasible: it indicates that the result is. An FPGA-based super simulator that is part of the trend of this article, any instrument. This is, Virtex-4 FPGA connected torus 64 Xilinx 2D 32 pieces of the visible CPU set of follows. Our machine is faster more than 100 times, practical Maxwell machine equivalent of the software is than it is possible to make a variety of Monte Carlo imitations. The GARCH (Generalized Autoregressive Conditional Heteroskedasticity) option is, in the context of the price model implementation, has been shown in this paper. To be excellent on the implementation of the equivalent of software that our FPGA-based function is running the same number of workstation clusters of computing terminal according to as the actual hardware features have shown the GARCH model.

A Parallel Iterative Finite Element Method for the Linear Elliptic Equations

By combining the two-grid discretization with the partition of unity method, a parallel iterative finite element method for the linear elliptic equations is proposed and investigated. Since the construction of the partition of unity is based on the coarse mesh triangulation, the computational domain of each subproblem can be divided automatically and the number of subproblems can be arbitrarily huge as the coarse mesh parameter H tends to zero. That means our method can be easily implemented in high performance supercomputers or cluster of workstations. Theoretical results based on a priori error estimation of the scheme are obtained, which indicate that our method can reach the optimal convergence orders within a few two-grid iterations. Numerical results are reported to assess the theoretical results.

Audio-visual preferences, perception, and use of water features in open-plan offices.

This paper examines the use of water features for masking irrelevant speech and improving the soundscape of open-plan offices. Two laboratory experiments were carried out, as well as acoustic simulations and field tests. Experiment 1 aimed to identify the preferred sound level of water sounds against irrelevant speech. Experiment 2 examined the audio-only and audio-visual preferences and perception of waterscapes. Acoustic simulations and field tests examined the impact of design factors. The results showed that, when played against a constant level of irrelevant speech of 48 dBA, people prefer to listen to water sounds of 42-48 dBA (45 dBA being best). These results and results from previous research suggest that water sounds work mainly as informational maskers rather than energetic maskers. Furthermore, the introduction of a water feature improved the perception of the sound environment, and adding visual stimuli improved perception by up to 2.5 times. Acoustic simulations indicated that features at each corner and one at the center (or a single feature with an array of speakers) can provide appropriate masking for a large open-plan office, whilst field tests showed that water sounds decrease the distraction and privacy distances significantly (clusters of workstations benefitting more than rows of workstations).

Workstation Cluster’s Hadoop Distributed File System Simulation and Modeling

Workstation Cluster’s Hadoop Distributed File System Simulation and Modeling

Strength Check of Aircraft Parts Based on Multi-GPU Clusters for Fast Calculation of Sparse Linear Equations

In order to improve the cost-effectiveness ratio, the next-generation vehicle needs to meet the requirements of reuse, while adopting a lighter structural weight, so it is necessary to realize the strength calculation and condition monitoring of key components in the digital twin. Most of the current monitoring methods are based on the characteristics of various data acquisition systems, but they need the support of a large number of flight data. The disadvantages of the above strategy can be avoided by reducing the structure of aircraft components to a finite element model and quickly checking the key components in the health management system. In order to solve the problem of fast calculation of the finite element model of the key components of the aircraft, a parallel algorithm and framework of large-scale sparse matrix preprocessing conjugate gradient method based on CUDA(Compute Unified Device Architecture) technology is proposed in the multi GPU(Graphics Processing Unit) workstation cluster environment. Once the sparse matrix is too large to be processed in a single workstation, this paper discusses how to realize the optimized data segmentation in the distributed multi-GPU computing environment. For the problem of iterative solution of matrix preprocessing, two preprocessing strategies of matrix bandwidth reduction parallelization and incomplete Cholesky decomposition are proposed, and asynchronous task concurrency and load balancing strategies are designed on the architecture. The calculation of some examples in the standard sparse matrix database shows that the algorithm and architecture proposed in this paper have the ability to solve large-scale sparse matrix quickly and efficiently, and can complete the fast strength verification of vehicle components.

A Fault Repair Method for Workstation Cluster Based on Probabilistic Model Checking

To analyze the component modules and maintenance unit modules in the workstation cluster, a fault repair method for workstation cluster based on probabilistic model checking is proposed. In the proposed method, the queue model is introduced when workstations are waiting for repairing, and different priorities are assigned according to the importance of the component functions. The formal model of the system is established by an extended continuous time Markov chain, the attributes of the system are described by continuous stochastic logic, and the fault repair module is verified by PRISM. The experimental results show that the proposed method can greatly reduce the time required for the maintenance process, and improve the maintenance efficiency and the fault tolerance of the system.

Fault tolerant adaptive parallel and distributed simulation through functional replication

This paper presents FT-GAIA, a software-based fault-tolerant parallel and distributed simulation middleware. FT-GAIA has being designed to reliably handle Parallel And Distributed Simulation (PADS) models, which are needed to properly simulate and analyze complex systems arising in any kind of scientific or engineering field. PADS takes advantage of multiple execution units run in multicore processors, cluster of workstations or HPC systems. However, large computing systems, such as HPC systems that include hundreds of thousands of computing nodes, have to handle frequent failures of some components. To cope with this issue, FT-GAIA transparently replicates simulation entities and distributes them on multiple execution nodes. This allows the simulation to tolerate crash-failures of computing nodes. Moreover, FT-GAIA offers some protection against Byzantine failures, since interaction messages among the simulated entities are replicated as well, so that the receiving entity can identify and discard corrupted messages. Results from an analytical model and from an experimental evaluation show that FT-GAIA provides a high degree of fault tolerance, at the cost of a moderate increase in the computational load of the execution units.

Efficient synthesis of robust models for stochastic systems

We describe a tool-supported method for the efficient synthesis of parametric continuous-time Markov chains (pCTMC) that correspond to robust designs of a system under development. The pCTMCs generated by our RObust DEsign Synthesis (RODES) method are resilient to changes in the system’s operational profile, satisfy strict reliability, performance and other quality constraints, and are Pareto-optimal or nearly Pareto-optimal with respect to a set of quality optimisation criteria. By integrating sensitivity analysis at designer-specified tolerance levels and Pareto optimality, RODES produces designs that are potentially slightly suboptimal in return for less sensitivity—an acceptable trade-off in engineering practice. We demonstrate the effectiveness of our method and the efficiency of its GPU-accelerated tool support across multiple application domains by using RODES to design a producer-consumer system, a replicated file system and a workstation cluster system.

Distributed simulation optimization and parameter exploration framework for the cloud

Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex real systems in different fields. Finding an optimal system design requires performing a large sweep over the parameter space in an organized way. Hence, the model optimization process is extremely demanding from a computational point of view, as it requires careful, time-consuming, complex orchestration of coordinated executions. In this paper, we present the design of SOF (Simulation Optimization and exploration Framework in the cloud), a framework which exploits the computing power of a cloud computational environment in order to carry out effective and efficient simulation optimization strategies. SOF offers several attractive features. Firstly, SOF requires “zero configuration”, as it does not require any additional software installed on the remote node; only standard Apache Hadoop and SSH access are sufficient. Secondly, SOF is transparent to the user, since the user is totally unaware that the system operates on a distributed environment. Finally, SOF is highly customizable and programmable, since it enables the running of different simulation optimization scenarios using diverse programming languages – provided that the hosting platform supports them – and different simulation toolkits, as developed by the modeler. The tool has been fully developed and is available on a public repository11SOF GitHub public repository, https://github.com/isislab-unisa/sof. under the terms of the open source Apache License. It has been tested and validated on several private platforms, such as a dedicated cluster of workstations, as well as on public platforms, including the Hortonworks Data Platform and Amazon Web Services Elastic MapReduce solution.

Parallel computations using a cluster of workstations to simulate elasticity problems

Computational physics has played important roles in real world problems. This paper is within the applied computational physics area. The aim of this study is to observe the performance of parallel computations using a cluster of workstations (COW) to simulate elasticity problems. Parallel computations with the COW configuration are conducted using the Message Passing Interface (MPI) standard. In parallel computations with COW, we consider five scenarios with twenty simulations. In addition to the execution time, efficiency is used to evaluate programming algorithm scenarios. Sequential and parallel programming performances are evaluated based on their execution time and efficiency. Results show that the one-dimensional elasticity equations are not appropriate to be solved in parallel with MPI_Send and MPI_Recv technique in the MPI standard, because the total amount of time to exchange data is considered more dominant compared with the total amount of time to conduct the basic elasticity computation.

Exhaustive Community Enumeration in Parallel

An algorithm to evaluate/count all the possible communities of a graph is presented. An associated unrank function is described. An implementation of an existing algorithm to evaluate all the possible partitions of a graph, based on an unrank function, is presented as well. Performance results of the parallelizations of these algorithms obtained on a shared memory machine, a cluster of workstations and a Graphical Processing Unit (GPU) are included.

Convectively unstable turbulent open channel flow with stable surface stratification

We use direct numerical simulations to examine fully developed turbulent open channel flow where the near wall region is unstably stratified and the outer boundary layer is stably stratified. The simulations are a model for flow in shallow turbid river channels with incident solar radiation. The aim is to determine under what conditions and by what mechanism the stably stratified layer is overturned. The flow is attained by applying a radiative heat flux at the free surface of the open channel. The absorption and transmission of the radiation follows the Beer--Lambert law with a constant absorption coefficient. We examine conditions where approximately 20% of the incident radiative heat flux penetrates through to the wall, releasing heat at the wall as a heat flux. The problem is specified by our buoyancy parameter which is analogous to the bulk Obukhov length scale. In the stable outer boundary layer we observe that the flux Richardson number reaches the limiting value, as was observed in the atmospheric boundary layer under sheared convective conditions. References S. M. Mitrovic, L. Hardwick, and F. Dorani. Use of flow management to mitigate cyanobacterial blooms in the Lower Darling River, Australia. J. Plankton Res. , 33(2):229–241, 2011. doi:10.1093/plankt/fbq094 J. R. Taylor, S. Sarkar, and V. Armenio. Large eddy simulation of stably stratified open channel flow. Phys. Fluids , 17(11):116602, 2005. doi:10.1063/1.2130747 M. Garcia-Villalba and J. C. del Alamo. Turbulence modification by stable stratification in channel flow. Phys. Fluids , 23(4):045104, 2011. doi:10.1063/1.3560359 N. Williamson, S. W Armfield, M. P. Kirkpatrick, and S Norris. A canonical model for stratified flow in estuaries and rivers. ANZIAM J. , 54:C88–C101, 2012. http://journal.austms.org.au/ojs/index.php/ANZIAMJ/article/view/6429 S. W. Armfield, S. E. Norris, P. Morgan, and R. Street. A parallel non-staggered Navier-Stokes solver implemented on a workstation cluster. In S. Armfield, P. Morgan, and K. Srinvas (Eds.) Computational Fluid Dynamics 2002: Proceedings of the Second International Conference on Computational Fluid Dynamics, Sydney, Australia, 15–19 July 2002 , pages 30–45. Springer, 2003. doi:10.1007/978-3-642-59334-5 R. J. Conzemius and E. Fedorovich. Dynamics of Sheared Convective Boundary Layer Entrainment. Part I: Methodological Background and Large-Eddy Simulations. J. Atmos. Sci. , 63(4):1151–1178, 2006. doi:10.1175/JAS3691.1

Analysis of Parallel Algorithms on SMP Node and Cluster of Workstations Using Parallel Programming Models with New Tile-based Method for Large Biological Datasets.

Sequence alignment is an important tool for describing the relationships between DNA sequences. Many sequence alignment algorithms exist, differing in efficiency, in their models of the sequences, and in the relationship between sequences. The focus of this study is to obtain an optimal alignment between two sequences of biological data, particularly DNA sequences. The algorithm is discussed with particular emphasis on time, speedup, and efficiency optimizations. Parallel programming presents a number of critical challenges to application developers. Today’s supercomputer often consists of clusters of SMP nodes. Programming paradigms such as OpenMP and MPI are used to write parallel codes for such architectures. However, the OpenMP programs cannot be scaled for more than a single SMP node. However, programs written in MPI can have more than single SMP nodes. But such a programming paradigm has an overhead of internode communication. In this work, we explore the tradeoffs between using OpenMP and MPI. We demonstrate that the communication overhead incurs significantly even in OpenMP loop execution and increases with the number of cores participating. We also demonstrate a communication model to approximate the overhead from communication in OpenMP loops. Our results are astonishing and interesting to a large variety of input data files. We have developed our own load balancing and cache optimization technique for message passing model. Our experimental results show that our own developed techniques give optimum performance of our parallel algorithm for various sizes of input parameter, such as sequence size and tile size, on a wide variety of multicore architectures.

Analyzing Google Earth application in a heterogeneous commodity cluster display wall

Nowadays, clusters of interconnected workstations have become a common solution for powering large composite displays, or “cluster display walls”, to visualize high resolution images. Our paper is focused on analyzing a specific cluster display wall developed by Google, named Liquid Galaxy, made up of heterogeneous commodity hardware with different degrees of heterogeneity, running master-slave (Google Earth) and client-server (Quake III Arena) multimedia applications. With this in mind, we define and test different scenarios, representing the behavior of many kinds of users. Our results show that the CPU, memory and network are good enough to execute the client-server application, while, depending on the user behavior, the external network constitutes the bottleneck of the system in Google Earth. So, the master-slave application has focused our attention. Likewise, in order to analyze the users’ point of view when interacting with Google Earth in the Liquid Galaxy, we define a new metric, named Visualization Rate (VR), which enables a relationship to be established between the user experience and the platform performance. In order to set the minimum acceptable value of the VR parameter according to users perception, we carried out different tests with real users. Then, this minimum threshold was compared with the VR value obtained from the automated benchmarking performed afterwards on clusters with different heterogeneity degrees. Finally, we analyzed the VR trend when the Liquid Galaxy is scaled from 3 up to 8 nodes in both the homogeneous and heterogeneous architectures to study the scalability of the system.

Design and evaluation of a parallel algorithm for inferring topic hierarchies

The rapid growth of information in the digital world especially on the web, calls for automated methods of organizing the digital information for convenient access and efficient information retrieval. Topic modeling is a branch of machine learning and probabilistic graphical modeling that helps in arranging the web pages according to their topical structure. The topic distribution over a set of documents (web pages) and the affinity of a document toward a specific topic can be revealed using topic modeling. Topic modeling algorithms are typically computationally expensive due to their iterative nature. Recent research efforts have attempted to parallelize specific topic models and are successful in their attempts. These parallel algorithms however have tightly-coupled parallel processes which require frequent synchronization and are also tightly coupled with the underlying topic model which is used for inferring the topic hierarchy. In this paper, we propose a parallel algorithm to infer topic hierarchies from a large scale document corpus. A key feature of the proposed algorithm is that it exploits coarse grained parallelism and the components running in parallel need not synchronize after every iteration, thus the algorithm lends itself to be implemented on a geographically dispersed set of processing elements interconnected through a network. The parallel algorithm realizes a speed up of 53.5 on a 32-node cluster of dual-core workstations and at the same time achieving approximately the same likelihood or predictive accuracy as that of the sequential algorithm, with respect to the performance of Information Retrieval tasks.

A parallel text document clustering algorithm based on neighbors

In this paper, we propose a new parallel algorithm for text document clustering based on the concept of neighbor (Guha et al. in Inf Syst 25(5):345---366, 2000). If two documents are similar enough, they are considered as neighbors of each other. The new algorithm is named parallel k-means based on neighbors (PKBN), and it is a parallel version of sequential k-means based on neighbors (SKBN) that we proposed in Luo et al. (Data Knowl Eng 68(11):1271---1288, 2009). PKBN fully exploits the data-parallelism of SKBN and adopts a new parallel pair-generating method to build the neighbor matrix. Our new parallel pair-generating method causes less communication overhead between processors than existing methods. PKBN is designed for message-passing multiprocessor systems and is implemented on a cluster of Linux workstations to analyze its performance. Our experimental results on real-life data sets demonstrate that PKBN is very efficient and has good scalability with respect to the number of processors and the size of data set.

AN EXPERIMENTAL EVALUATION OF PERFORMANCE OF A HADOOP CLUSTER ON REPLICA MANAGEMENT

Hadoop is an open source implementation of the MapReduce Framework in the realm of distributed processing. A Hadoop cluster is a unique type of computational cluster designed for storing and analyzing large data sets across cluster of workstations. To handle massive scale data, Hadoop exploits the Hadoop Distributed File System termed as HDFS. The HDFS similar to most distributed file systems share a familiar problem on data sharing and availability among compute nodes, often which leads to decrease in performance. This paper is an experimental evaluation of Hadoop's computing performance which is made by designing a rack aware cluster that utilizes the Hadoop's default block placement policy to improve data availability. Additionally, an adaptive data replication scheme that relies on access count prediction using Langrange's interpolation is adapted to fit the scenario. To prove, experiments were conducted on a rack aware cluster setup which significantly reduced the task completion time, but once the volume of the data being processed increases there is a considerable cutback in computational speeds due to update cost. Further the threshold level for balance between the update cost and replication factor is identified and presented graphically.